AGV positioning and docking method

Abstract

The application provides an AGV positioning and docking method, reflective plates are arranged on both sides in front of a machine table, the length of the machine table is c, the width of the machine table is d, a two-dimensional code is arranged at the center position of the bottom of the machine table, the distance of the center point of the two-dimensional code is directly determined through different coordinate system conversion to realize positioning, the middle position of the reflective plate is driven into the bottom of the machine table to realize code scanning and docking, so that the AGV can accurately drive to the bottom of the machine table to realize code scanning and complete docking, and the method is simple and reliable.

Description

Technical Field

[0001] The present invention relates to the technical field of robot navigation, and in particular to an AGV (Automated Guided Vehicle) positioning and docking method. Background Art

[0002] In recent years, with the continuous improvement of production technology, most modern intelligent manufacturing workshops are equipped with intelligent mobile robots, and AGV carts are one of the intelligent mobile robots. They locate and navigate to dock with other logistics equipment to achieve loading and unloading and transportation of materials. In this process, compared with the traditional rail navigation docking method, the positioning and navigation technology based on laser radar is the key technology for AGV carts to navigate and complete docking in indoor workshops. The navigation method based on laser radar has high accuracy.

[0003] For example, Chinese patent application number 201710571639.9, published on September 8, 2017, discloses a laser positioning and navigation method based on dual reflectors. The method requires at least two reflectors to be detected, but in actual application, multiple reflectors may be detected simultaneously. By designing data fusion processing, the complex domain information of the reflectors can be maximized, thereby further improving accuracy.

[0004] In the workshop, when the AGV needs to be controlled to transport materials under the machine, the AGV needs to accurately find the positioning point before it can drive to the bottom of the machine to dock. However, the double reflectors in the above literature only allow the AGV to drive in front of the reflector of the machine. At the same time, if the AGV is close to the reflector in front of the machine, the data of the scanned reflector may be uncertain and deviated during the matching process of the reflector. However, if a QR code is used for navigation under the machine, since the QR code needs to be preset at the bottom of the machine, if the machine is displaced during the movement of the AGV, the movement to the position without the QR code will make it impossible to navigate to the bottom of the machine after the movement, thereby failing to ensure that the AGV accurately drives to the bottom of the machine to scan the code for docking, and then complete the loading and unloading and transportation of materials. Summary of the Invention

[0005] The purpose of the present invention is to provide an AGV trolley positioning and docking method, which enables the AGV trolley to accurately drive to the bottom of the machine to scan the code and complete the docking. The method is simple and reliable.

[0006] To achieve the above-mentioned purpose, the present invention provides an AGV positioning and docking method, in which reflectors are provided on both sides of the front of the machine, the length of the machine is preset to be c, the width of the machine is preset to be d, and a QR code is set at the center of the bottom of the machine. Positioning is achieved by directly determining the distance to the center point of the QR code through different coordinate system conversions. The AGV is driven from the middle position of the reflector to the bottom of the machine to scan the code for docking.

[0007] The following steps are involved:

[0008] S1: Control the AGV to enter the reflector scanning area in front of the machine and switch to the reflector positioning mode, and then start matching the reflector outline;

[0009] S1.1, scan the point cloud data in the reflector area by the laser radar installed on the AGV, preset the point cloud intensity threshold T, weighting coefficients α, β, γ,

[0010] S1.2. Collect the scanned point cloud data, and then filter the point cloud data according to the preset intensity threshold T.

[0011] S1.3. Determine the error between the contour length of the filtered point cloud data and the length of the reflector, determine the center-to-center spacing error between the filtered point cloud data and the reflector, and determine the relative intensity error between the filtered point cloud data and the reflector. Use the evaluation function cost to determine whether the reflector contour matches; the evaluation function cost is a weighted coefficient and a weighted sum of the length error, center-to-center spacing error, and relative intensity error.

[0012] S2. Obtain the pose matrix of the reflector in the lidar coordinate system through the coordinates of the point cloud data , then according to the transformation matrix T of the laser radar in the odometer coordinate system b Get the pose matrix T of the reflector in the odometer coordinate system g , according to the transformation matrix T of the origin of the odometer coordinate system in the AGV vehicle coordinate system r Get the pose matrix T of the reflector in the AGV coordinate system L ;

[0013] S3, according to the pose matrix T of the reflector in the AGV coordinate system L Determine whether the AGV is on the straight path where the center position between the reflectors is located. If there is a deviation error, correct the deviation error and return the AGV to the straight path where the center position between the reflectors is located.

[0014] S4. Establish a reflector coordinate system with the center of the reflector on one side of the machine as the origin. Then, determine the position coordinates of the center point of the QR code in the reflector coordinate system as (x1, y1, z1) based on the length and width of the machine.

[0015] S5. Preset the transformation matrix between the reflector coordinate system and the odometer coordinate system as T1. The position matrix T3 of the center point of the QR code in the odometer coordinate system can be obtained through the transformation matrix T1. Preset the transformation matrix between the AGV coordinate system and the odometer coordinate system as T2. By combining the position matrix T3 with the transformation matrix T2, the position matrix T4 of the center point of the QR code in the AGV coordinate system can be obtained.

[0016] S6. Calculate the straight-line distance between the center point of the QR code and the AGV through the position matrix T4, and then control the AGV to drive straight to the scanning area at the center of the bottom of the machine to achieve scanning docking.

[0017] The above settings scan the point cloud data in the scanning area in front of the machine through the laser radar, and preset the point cloud intensity threshold T. This makes it easier to compare the preset point cloud intensity threshold T with the intensity of the scanned point cloud data, thereby screening out the point cloud data that matches the reflector, eliminating other interference factors, and improving accuracy; then calculate the coordinates of the filtered point cloud data (lx i ,ly i ) Get the pose matrix of the reflector in the lidar coordinate system , and then according to the transformation matrix T of the laser radar in the odometer coordinate system b , so that the pose matrix T of the reflector in the odometer coordinate system can be obtained g At the same time, the transformation matrix T of the origin of the odometer coordinate system in the AGV car coordinate system r , so that the pose matrix T of the reflector in the AGV coordinate system can be obtained L , so that the reflector can be used to determine the pose matrix T in the AGV coordinate system LDetermine whether the AGV is on the straight path where the center position between the reflectors is located, and ensure that the AGV is on the straight path where the center position between the reflectors is located; by setting the transformation matrix between the reflector coordinate system and the odometer coordinate system to T1, and setting the transformation matrix between the AGV coordinate system and the odometer coordinate system to T2, the position matrix T3 of the center point of the QR code in the odometer coordinate system can be obtained through the transformation matrix T1, so that the position matrix T3 and the transformation matrix T2 can be combined to calculate the position matrix T4 of the center point of the QR code in the AGV coordinate system, and then calculate the straight-line distance between the center point of the QR code and the AGV. Then, the AGV is controlled to drive straight to the scanning area at the center position of the bottom of the machine to realize scanning docking and realize material loading, unloading and handling. For example, during the movement of the AGV, the machine moves, and the center of the QR code at the bottom is the center of the AGV. Then, according to the center of the AGV, it is mapped to the reflector coordinate system, and then the AGV movement path is obtained after matrix transformation, so that the AGV moves to the machine position for docking.

[0018] Furthermore, the err_length function is used to calculate the error between the contour length of the filtered point cloud data and the length of the reflector, the err_dist function is used to calculate the center distance error between the filtered point cloud data and the reflector, and the cost_intensity function is used to calculate the relative intensity error between the filtered point cloud data and the reflector. The evaluation function cost is used to determine whether the reflector contour matches.

[0019] With the above settings, the length error, center distance error, and relative intensity error between the filtered point cloud data and the reflector are calculated through the err_length function, err_dist function, and cost_intensity function, respectively. Then, the evaluation function cost can be obtained and used to determine whether the reflector contour matches. In this way, the AGV can complete the contour matching of the scanned reflector.

[0020] Furthermore, the step S1.2 further includes:

[0021] The intensity threshold T0 of the scanned point cloud data is compared with the preset intensity threshold T. If T0 is within the range of [0.6*T, 0.7*T], it is judged as the point cloud data of the reflector; otherwise, it is judged as other interference point cloud data.

[0022] The above settings can filter out the point cloud data that matches the reflector from all the scanned point cloud data.

[0023] Furthermore, the step S1.3 further includes:

[0024] The formula of the evaluation function cost is as follows:

[0025] cost=α* err_length+ β*err_dist+ γ*cost_intensity(1).

[0026] The above settings make it easy to determine whether the reflector profile matches through the evaluation function cost.

[0027] Furthermore, the step S2 also includes calculating the coordinates of the filtered point cloud data (lx i ,ly i ), the specific steps are as follows:

[0028] S2.1. Obtain the distance l between the laser radar and the reflector and the scanning angle z from the laser radar scan.

[0029] S2.2, then use the distance l and the scanning angle z to calculate lx respectively i ,ly i ,

[0030] lx i =l*cos(z)(2),

[0031] ly i =l*sin(z)(3),

[0032] S2.3. Calculate the pose matrix of the reflector in the LiDAR coordinate system as follows,

[0033] (4),

[0034] Calculate the pose matrix T of the reflector in the odometry coordinate system g as follows,

[0035] (5),

[0036] Calculate the pose matrix T of the reflector in the AGV coordinate system L as follows,

[0037] T L =T g *T r (6).

[0038] The above settings can calculate the filtered point cloud coordinates by the distance l between the lidar and the reflector and the scanning angle z; the pose matrix of the reflector in the lidar coordinate system can be obtained by the coordinates of the point cloud data. , and then calculate the pose matrix T of the reflector in the odometer coordinate system gFinally, we get the pose matrix T of the reflector in the AGV coordinate system. L .

[0039] Furthermore, the step S3 further includes steps S3.1-S3.2:

[0040] S3.1. If the AGV deviates from the straight path where the center position between the reflectors is located, then according to the pose matrix T of the reflectors in the AGV coordinate system L Calculate the deviation error and then control the AGV to return to the straight path.

[0041] S3.1.1, take the straight line path as the y-axis and the horizontal line where the AGV is located as the x-axis, and then establish a rectangular coordinate system.

[0042] S3.1.2, according to the pose matrix T of the reflector in the AGV coordinate system L Determine the distance y0 that the AGV deviates from the y-axis, then select the deviation angle b0, and use the geometric relationship to calculate the distance S that the AGV travels to the y-axis as follows:

[0043] (7),

[0044] S3.1.3. Control the AGV to travel a distance S at a speed v0 along a deviation angle b0 and stop. Then control the AGV to rotate counterclockwise by (90-b0)° to correct the deviation error and return to the straight path.

[0045] S3.2: If the AGV does not deviate from the straight path where the center position between the reflectors is located, proceed to step S4.

[0046] The above settings can correct the deviation error of the AGV car from the straight path where the center position between the reflectors is located after matching the reflector contour, so that the AGV car can accurately return to the straight path.

[0047] Furthermore, the step S5 further includes steps S5.1-S5.2:

[0048] S5.1. Calculate the position matrix T3 of the center point of the QR code in the odometry coordinate system as follows:

[0049] (8);

[0050] S5.2. By combining the position matrix T3 with the transformation matrix T2, the position matrix T4 of the center point of the QR code in the AGV coordinate system is calculated as follows:

[0051] T4= T2* T3 (9).

[0052] The above settings can calculate the position of the center point of the QR code in the AGV coordinate system.

[0053] Furthermore, step S6 further includes the following:

[0054] If the AGV is within the docking error range, the AGV completes the docking, thereby realizing the loading, unloading and transportation of materials; otherwise, the docking is performed again.

[0055] The above settings can enable the AGV to complete docking within the docking error range, thereby realizing the loading, unloading and transportation of materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flow chart of the working method of the present invention.

[0057] Figure 2 This is a geometric diagram of the calculation during the deviation error correction process of the AGV in the present invention. DETAILED DESCRIPTION

[0058] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0059] like Figure 1 As shown in the figure, a positioning and docking method for an AGV is provided. Reflectors are provided on both sides of the front of the machine. The length of the machine is preset to be c, the width of the machine is preset to be d, and a QR code is provided at the center of the bottom of the machine. Positioning is achieved by directly determining the distance to the center point of the QR code through different coordinate system conversions. The vehicle drives from the middle position of the reflector to the bottom of the machine to scan the code for docking.

[0060] The specific embodiments and steps are as follows:

[0061] S1, control the AGV to enter the reflector scanning area in front of the machine and switch to the reflector positioning mode, and then start matching the reflector outline.

[0062] S1.1, scan the point cloud data in the reflector area by the laser radar installed on the AGV, preset the point cloud intensity threshold T, weighting coefficients α, β, γ,

[0063] S1.2. Collect the scanned point cloud data, and then filter the point cloud data based on a preset intensity threshold T. In this embodiment, the intensity threshold T0 of the scanned point cloud data is compared with the preset intensity threshold T. If T0 is within the range of [0.6*T, 0.7*T], the point cloud data is determined to be the reflector; otherwise, it is determined to be other interference point cloud data.

[0064] S1.3. Calculations are performed using functions in the INS / GPC navigation system. The err_length function is used to calculate the error between the contour length of the filtered point cloud data and the length of the reflector. Specifically, the err_length function determines the contour length of the point cloud data using the positioning system and then compares it with the length information of the preset reflector to calculate the length error. The err_dist function is used to calculate the center distance error between the filtered point cloud data and the reflector. The err_dist function determines the error by the difference in distance between the center of the point cloud data determined by the positioning system and the center of the preset reflector. The cost_intensity function is used to calculate the relative intensity error between the filtered point cloud data and the reflector. The cost_intensity function determines the relative intensity error by comparing the intensity information of the point cloud data in the image with the intensity information of the preset reflector. The evaluation function cost is used to determine whether the reflector contour matches. In this embodiment, the err_length function, err_dist function, and cost_intensity function are all existing functions and will not be repeated here. The formula for the evaluation function cost is as follows:

[0065] cost=α* err_length+ β*err_dist+ γ*cost_intensity(1),

[0066] If the reflector profile is matched successfully, proceed to step S2;

[0067] If the reflector profile is not matched successfully, the process returns to step S1.

[0068] S2. Obtain the pose matrix of the reflector in the lidar coordinate system through the coordinates of the point cloud data , then according to the transformation matrix T of the laser radar in the odometer coordinate system b Get the pose matrix T of the reflector in the odometer coordinate system g , according to the transformation matrix T of the origin of the odometer coordinate system in the AGV vehicle coordinate system r Get the pose matrix T of the reflector in the AGV coordinate system L ,

[0069] S2.1. Obtain the distance l between the laser radar and the reflector and the scanning angle z during the laser radar scanning in step S1.1;

[0070] S2.2, then use the distance l and the scanning angle z to calculate lx respectively i ,ly i ,

[0071] lx i =l*cos(z)(2),

[0072] lyi =l*sin(z)(3);

[0073] S2.3, then determine the body angle through the posture sensor, and then obtain the posture matrix of the reflector in the lidar coordinate system through the coordinates of the point cloud data and the body angle value The pose matrix includes the horizontal and vertical coordinates of the two-dimensional space and the body angle information. In this embodiment, the pose matrix of the reflector in the laser radar coordinate system is The calculation formula is as follows: Since the mobile robot only moves in the plane, and the plane is the surface where the XY axis is located, and the mobile robot will not climb;

[0074] (4),

[0075] Calculate the pose matrix T of the reflector in the odometry coordinate system g as follows,

[0076] (5),

[0077] Calculate the pose matrix T of the reflector in the AGV coordinate system L as follows,

[0078] T L =T g *T r (6).

[0079] S3, according to the pose matrix T of the reflector in the AGV coordinate system L Determine whether the AGV is on the straight path at the center between the reflectors.

[0080] S3.1. If the AGV deviates from the straight path where the center position between the reflectors is located, then according to the pose matrix T of the reflectors in the AGV coordinate system L Calculate the deviation error and then control the AGV to return to the straight path.

[0081] S3.1.1. Use the straight path as the y-axis and the horizontal line where the AGV is located as the x-axis, and then establish a rectangular coordinate system;

[0082] S3.1.2, according to the pose matrix T of the reflector in the AGV coordinate system L Determine the distance y0 that the AGV deviates from the y-axis, then select the deviation angle b0, and use the geometric relationship to calculate the distance S that the AGV travels to the y-axis. In this embodiment, the deviation angle b0 is set to 20° as follows:

[0083] (7);

[0084] S3.1.3. Control the AGV to travel a distance S at a speed v0 along a deviation angle b0 and stop. Then control the AGV to rotate (90-b0)° in the direction of the straight path, counterclockwise in this embodiment, so that the AGV returns to the straight path after correcting the deviation error.

[0085] S3.2: If the AGV does not deviate from the straight path where the center position between the reflectors is located, proceed to step S4.

[0086] S4. Establish a reflector coordinate system with the center of the reflector on one side of the machine as the origin. Then, determine the position coordinates of the center point of the QR code in the reflector coordinate system as (x1, y1, z1) based on the length and width of the machine.

[0087] S5. The transformation matrix between the reflector coordinate system and the odometer coordinate system is preset as T1. The position matrix T3 of the center point of the QR code in the odometer coordinate system can be obtained through the transformation matrix T1. The transformation matrix between the AGV coordinate system and the odometer coordinate system is preset as T2. The position matrix T3 is combined with the transformation matrix T2 to obtain the position matrix T4 of the center point of the QR code in the AGV coordinate system.

[0088] S5.1. Calculate the position matrix T3 of the center point of the QR code in the odometry coordinate system as follows:

[0089] (8);

[0090] S5.2. By combining the position matrix T3 with the transformation matrix T2, the position matrix T4 of the center point of the QR code in the AGV coordinate system is calculated as follows:

[0091] T4= T2* T3 (9).

[0092] S6. Calculate the straight-line distance L0 between the center point of the QR code and the AGV through the position matrix T4, and then control the AGV to drive straight to the scanning area at the center of the bottom of the machine to achieve scanning docking.

[0093] The working principle of the present invention is as follows: the AGV is controlled to enter the reflector area in front of the machine, and then the reflector positioning mode is switched to scan the reflector. Then, the point cloud data obtained by the scan is filtered according to the preset intensity threshold T, and the filtered point cloud data is judged as the point cloud data of the reflector. Then, the error between the contour length of the filtered point cloud data and the length of the reflector is calculated by the err_length function, the error between the center point of the filtered point cloud data and the reflector is calculated by the err_dist function, and the relative intensity error between the filtered point cloud data and the reflector is calculated by the cost_intensity function. After the reflector contour is judged to be successfully matched using the evaluation function cost, the coordinates of the filtered point cloud data (lx i ,ly i ) and obtain the pose matrix of the reflector in the laser radar coordinate system , and then according to the transformation matrix T of the laser radar in the odometer coordinate system b Get the pose matrix T of the reflector in the odometer coordinate system g , according to the transformation matrix T of the origin of the odometer coordinate system in the AGV vehicle coordinate system r Get the pose matrix T of the reflector in the AGV coordinate system L At the same time, the pose matrix T of the reflector in the AGV coordinate system L Determine whether the AGV is on the straight path where the center position between the reflectors is located, so that the AGV can correct the deviation error from the straight path where the center position between the reflectors is located, so that the AGV can accurately return to the straight path, and then use the center of the reflector on one side of the machine as the origin to establish the reflector coordinate system and determine the position of the QR code center point in the reflector coordinate system. Then, calculate the position matrix T3 of the QR code center point in the odometer coordinate system through the transformation matrix T1 of the reflector coordinate system and the odometer coordinate system. At the same time, combine the transformation matrix T2 of the AGV coordinate system and the odometer coordinate system to obtain the position matrix T4 of the QR code center point in the AGV coordinate system, calculate the straight-line distance between the QR code center point and the AGV, and then control the AGV to drive straight to the scanning area at the center position of the bottom of the machine to realize scanning docking.

Claims

1. A method for positioning and docking an AGV, wherein reflectors are provided on both sides of the front of a platform, the platform length is preset to c, the platform width is preset to d, a QR code is provided at the center of the bottom of the platform, and positioning is achieved by directly determining the distance from the center of the QR code through different coordinate system conversions. The vehicle is driven from the center of the reflector to the bottom of the platform, and the code is scanned to achieve docking. The method is characterized by: The following steps are involved: S1: Control the AGV to enter the reflector scanning area in front of the machine and switch to the reflector positioning mode, and then start matching the reflector outline; S1.

1. Scan the point cloud data in the reflector area using a laser radar installed on the AGV, preset a point cloud intensity threshold T, and weighting coefficients α, β, and γ. S1.

2. Collect the scanned point cloud data, and then filter the point cloud data according to the preset intensity threshold T. S1.

3. Determine the error between the contour length of the filtered point cloud data and the length of the reflector, the center distance error between the filtered point cloud data and the reflector, and the relative intensity error between the filtered point cloud data and the reflector. Use the evaluation function cost to determine whether the reflector contour matches; the evaluation function cost is the weighted sum of the weighting coefficient and the length error, the center distance error, and the relative intensity error. S2. Obtain the pose matrix of the reflector in the LiDAR coordinate system through the coordinates of the point cloud data , then according to the transformation matrix T of the laser radar in the odometer coordinate system b Get the pose matrix T of the reflector in the odometer coordinate system g , according to the transformation matrix T of the origin of the odometer coordinate system in the AGV vehicle coordinate system r Get the pose matrix T of the reflector in the AGV coordinate system L ; S3, according to the pose matrix T of the reflector in the AGV coordinate system L Determine whether the AGV is on the straight path where the center position between the reflectors is located. If there is a deviation error, correct the deviation error and return the AGV to the straight path where the center position between the reflectors is located. S4. Establish a reflector coordinate system with the center of the reflector on one side of the machine as the origin. Then, determine the position coordinates of the center point of the QR code in the reflector coordinate system as (x1, y1, z1) based on the length and width of the machine. S5. Preset the transformation matrix between the reflector coordinate system and the odometer coordinate system as T1. The position matrix T3 of the center point of the QR code in the odometer coordinate system can be obtained through the transformation matrix T1. Preset the transformation matrix between the AGV coordinate system and the odometer coordinate system as T2. By combining the position matrix T3 with the transformation matrix T2, the position matrix T4 of the center point of the QR code in the AGV coordinate system can be obtained. S6. Calculate the straight-line distance between the center point of the QR code and the AGV through the position matrix T4, and then control the AGV to drive straight to the scanning area at the center of the bottom of the machine to achieve scanning docking.

2. The AGV positioning and docking method according to claim 1, characterized in that: The err_length function is used to calculate the error between the contour length of the filtered point cloud data and the length of the reflector. The err_dist function is used to calculate the center distance error between the filtered point cloud data and the reflector. The cost_intensity function is used to calculate the relative intensity error between the filtered point cloud data and the reflector. The evaluation function cost is used to determine whether the reflector contour matches.

3. The AGV positioning and docking method according to claim 1, characterized in that: The step S1.2 further includes: The intensity threshold T0 of the scanned point cloud data is compared with the preset intensity threshold T. If T0 is within the range of [0.6*T, 0.7*T], it is judged as the point cloud data of the reflector; otherwise, it is judged as other interference point cloud data.

4. The AGV positioning and docking method according to claim 2, characterized in that: The step S1.3 further includes: The formula of the evaluation function cost is as follows: cost=α* err_length+ β*err_dist+ γ*cost_intensity(1).

5. The AGV positioning and docking method according to claim 1, characterized in that: The step S2 also includes calculating the coordinates of the filtered point cloud data (lx i ,ly i ), the specific steps are as follows: S2.

1. Obtain the distance l between the laser radar and the reflector and the scanning angle z from the laser radar scan. S2.2, then use the distance l and the scanning angle z to calculate lx respectively i ,ly i , lx i =l*cos(z)(2), ly i =l*sin(z)(3), S2.

3. Calculate the pose matrix of the reflector in the LiDAR coordinate system as follows, （4）， Calculate the pose matrix T of the reflector in the odometry coordinate system g as follows, （5）， Calculate the pose matrix T of the reflector in the mobile car coordinate system L as follows, T L =T g *T r (6)。 6. The AGV positioning and docking method according to claim 1, characterized in that: The step S3 further includes steps S3.1-S3.2: S3.

1. If the mobile car deviates from the straight path where the center position between the reflectors is located, then according to the pose matrix T of the reflectors in the mobile car coordinate system L Calculate the deviation error and then control the mobile car to return to the straight path. S3.1.

1. Use the straight path as the y-axis and the horizontal line where the mobile car is located as the x-axis, and then establish a rectangular coordinate system. S3.1.2, according to the pose matrix T of the reflector in the coordinate system of the mobile car L Determine the distance y0 that the mobile car deviates from the y-axis, then select the deviation angle b0, and use the geometric relationship to calculate the distance S that the mobile car travels to the y-axis as follows: （7）， S3.1.

3. Control the mobile car to travel at a speed v0 along the deviation angle b0 and a distance S to stop. Then control the mobile car to rotate counterclockwise by (90-b0) degrees to correct the deviation error and return to the straight path. S3.

2. If the moving carriage does not deviate from the straight path where the center position between the reflectors is located, the moving carriage is controlled to move along the straight path where the center position between the reflectors is located and enter the bottom of the machine.

7. The AGV positioning and docking method according to claim 1, characterized in that: The step S5 further includes steps S5.1-S5.2: S5.

1. Calculate the position matrix T3 of the center point of the QR code in the odometry coordinate system as follows: （8）； S5.

2. By combining the position matrix T3 with the transformation matrix T2, the position matrix T4 of the center point of the QR code in the AGV coordinate system is calculated as follows: T4= T2* T3 (9).

8. The AGV positioning and docking method according to claim 1, characterized in that: In step S6, Includes the following: If the AGV is within the docking error range, the AGV completes the docking, thereby realizing the loading, unloading and transportation of materials; otherwise, the docking is performed again.

Images